32×32 is slowly becoming the new 16×16.

16×16 was born

First of all, let’s go down the memory lane.

“Icons” were one of the most important part of GUIs (Graphical User Interface), and surely one that sets it apart from CLI (Command Line Interface). In the very beginning, there was only 32×32. The earliest graphical operating systems, e.g. Mac System 1.0, spotted only 32×32 icons.

Read the rest of this entry »

In chronological order from left to right: Office 12 alpha, Office 2007 (12 final), Office 14 alpha

The story of the Ribbon

It has been a long time since Jensen Harris had first posted his design processes for Office 2007. I enjoyed reading it a lot, as he explained every step in detail, from how the design had evolved to how feedbacks were made into use. He had not updated his blog ever since, and then the next version of Office quietly kicks into development.

One of his major contributions was the “Ribbon”, as a way to consolidate all toolbars and to prevent a degrading experience due to accumulating toolbars and moving buttons around unknowingly. Basically, menus and toolbars are melt into one new paradigm, where buttons are not sized, hidden and positioned the same simply for organization and consistency, but also grouped, sized and visualized by importance and the power of suggestion through visual effects.

The Ribbon has ever since become the talking point of most reviewers and users, and soon Office 2007 becomes the example of how an old software can be revolutionized through UI. It has become so successful that the Windows division decided to bring the Ribbon interface over to its apps, for example, Paint, WordPad and Windows Live Movie Maker. (As a note, the default UI of Trillian Astra is also influenced by the Ribbon, but that’s another blog post.)

Throwing a Fitt!

However, the problem comes when Windows 7 begins adopting the Ribbon. As if Mr. Harris had retired from user interface design, the experience begins to disintegrate through its newer iterations, perhaps through decisions of other designers (who think they know best).

Though the changes are subtle, they are also disturbing and they show the lack of understanding of some very basic concepts:

1. Lack of visual distinctions

Let’s begin with the small visual details. While replacing boxes with separators considerably cleaned the interface up, the original boxes are not that suffocating to justify its removal. Instead, without the boxes, it is much harder to tell what functions belong together.

Top: Clean, distinctive grouping of functions (Office 2007)
Bottom: A clusterfsck of vertical noises (Office 14 Alpha)

Besides, not only the separators cannot do the job of the boxes, they contribute even more visual noises than the boxes. In attempt to make the separations clear, the designers put a darker line for the separators than the boxes. The result is that the lines have gotten so dark it is now fighting with the icons.

2. Where is the title bar again?

 
Top: I’m editing "Document" (WordPad in Windows Vista)
Middle: So I’m editing… uh… "Document" (WordPad in Windows 7)
Bottom: Wait, what am I editing? (Word 2007)

Strangely enough, the title bar in Office 2007 is centered, while the title bar in Windows 7 is aligned left. And Office 14? Centered. The title bar is important because there is no other way a user can tell what document they are editing. When one glances at the usual position, i.e. top left, one will find that the title is not there, but instead, jumping around the center, depending on how many contextual tabs you had opened.

3. Realignment of the Orb

Besides the minute changes above, the biggest offender is that they moved Office menu button. While the Office orb is not the most elegant solution, moving the menu button down with the tabs is worse. The reason?

Yes, Fitts’s Law. Of course.

In the original Office 2007 design, the Office menu occupied the whole top-left corner. Comparing the new design, it had two obvious benefits:

1. The orb is twice bigger. Therefore, when the window is floating, it is twice easier to click. And besides the physical benefit, a bigger button is distinctive. It is like putting a watermelon between apples, versus putting an orange between apples. While you can still point out that there is an orange upon closer look, a watermelon is a lot more obvious.

2. The orb is infinitely bigger when the window is maximized. Though not visually represented, it was coded so that you can simply push the cursor to the top left of the screen to click the button.

The functions included in the menu justified the placement of the orb. It was a very good design for it reduced a previously menu-ridden Office into just one menu – an impressive feat indeed.

To add salt to injury, its replacement in Office 14 alpha is nowhere as good. Back in the topleft corner is the useless window menu, which consists of ‘move’, ‘minimize’ and other boring functions that are much better performed by the mouse.

And worse, the Office orb is squeezed to look like… a tab. WHY? A menu looks NOTHING like the rest of the tabs in the ribbon. This is very bad design because it breaks user expectations, especially when users are nowhere familiar with the Ribbon yet. While users just learned that "all tabs switched the whole top toolbar away", now they would think that "some tabs switched the whole top toolbar away; but some tabs would blow me away with a full screen menu". That’s BAD.

Above: That explosive Office 14 menu I was talking about. It’s so big the Windows Start Menu looks like a midget.

Nevertheless, they picked the right applications to convert to the Ribbon interface. I can see that the ‘editing’ software, i.e. WordPad, Word, Paint, Movie Maker, etc., will pick up the Ribbon, while the ‘viewing’ software like Photo Gallery or Media Player will be left in the bland icon-less, text-only toolbar-ed interface.

It’s ironic how the Ribbon effectively prevented users from degrading the experience over time, but it cannot prevent other designers or programmers from disintegrating it from its very own self! I guess corporations are amnesiac by their nature. What’s the point of patenting and protecting the design, when they are the first people to wreck it?

Please come back, Mr Jensen Harris!

Not exactly a feat in design, but a feat in code optimization.

This skin was made as a proof of concept long time ago, when I designed the Custom Windows feature. As skins get more and more complex (Trillian Cordonata is 700+KB of hand code), the goal is to create a new subsystem that allows creation of skin in the least amount of code possible.

The solution is to have a base skin that encompasses everything that needs to be skinned, and to allow major parts of the skin (e.g. Contact List, Chat Windows) to be swap-able. It is unprecedented in any skinning engines in the world, as Trillian slowly becomes highly intricate and increasingly complex.

Originally based on Stixe, Custom Windows allow creation of skins that is semi-dependent on the base skin. The result is the ability for users to mix and match bits and parts of a skin, and the ability for skinners to easily create a skin as a weekend project.

“A Good Skin” contains around 20 lines of code, and only take an hour to bash out the graphics and codes.

Comparing the months and years spent on Trillian Cordonata, this is quite a feat.

You can download the skin here for a test drive.

Another day, another incremental build forward for Trillian Astra. As the new build 77 features a new Chat Room Browser UI design, I designed a few new icons to accompany it.

The main problem of a “Chat Room” icon is that there is hardly ever a general icon designed for it. A search on Google Images for idea reveals nothing particularly interesting:

Most designs retreats to the method used by Chinese characters as concepts get more and more complicated, i.e. 3 chat bubbles imply that it is not 1-to-1 conversation, thus a chat room or a conference:

This design would work, but, until the concept of “Chat Room Groups” come up in the Yahoo! Chat Browser. And very soon we would find ourselves drowning in the vortex of the most complicated Chinese character if we put forward such logic:

The solution is to replace the icon with a new concept. In this case, I drew a little House with a chat bubble blowing through its chimney to represent a Chat Room. As users learn that a house represents a Chat Room, they will also understand that a few houses represents a group of Chat Rooms. (Let’s see how many other chat clients and chat sites would rip this idea off.)

I know, this is pretty minute, but it saves you from a bubble bath in the Chat Room Browser.

This is a 3-part series on porting Trillian across platforms, particularly Mac OS X. It would make more sense if you start reading from Part 1. Thanks for your interest!

Coherence across platforms

Meanwhile, the second question is the coherence of the product between different platforms. If the user switches between Web and Mac OS X often, it would be nice if their experiences are around the same. Our web version is a very delicate replica of the Windows version, so that transition between Web and Windows is very smooth. Users don’t need to re-learn how to use again, as it is conceived as one single product. It would be not convenient if in one OS the toolbar is at the top, and in another the toolbar is at the bottom.

In order to provide a familiar experience for all Trillian Astra users across the platform, we use the same skin through out. While we made a lot of changes to accommodate the ease of use in Mac OS X, the form of the contact list window remained the same – with its signature curved bottom housing a variety of controls – and it works well for the identity of the instant messenger as well.

The OS X version of Cordonata has an overall flatter “lacquered and polished steel” appearance that is based on Leopard. I am not completely sure about the darkness of the window yet, but it seems most windows in Leopard are very gray. We have also flattened the bottom buttons to follow the new toolbar buttons in Leopard.

Font choice varies across the platforms. It makes sense because certain system fonts do not exist on other systems. Default font of respective operating systems are used as a result, while in Flash we will try to adapt to both.

Other Trillian-centric behaviors are kept in spots where familiarity between Trillian products weighs more beneficial than familiarity between Trillian and OS X, and where there were no precedents in Mac OS X. For example, we use the same toolbar in both Windows and Mac to showcase all the IM services the user is using.

So this concludes my 3-part series on porting the design of Trillian to Mac OS X. Please feel free to send me comments on what you would expect on the designs of Trillian. Thanks!

Coherence with Mac OS X

There were two stages in the design of the windows. While critics may say it is unnecessary to design anything new with the use of HIG (Human Interface Guidelines) from Apple, it is in no doubt instant messaging requires a large variety of new controls.

For instance, Apple had to introduce chat bubbles in iChat to distinguish individual messages in a conversation. In the contact list, there are no standards between how contacts are laid out. Some have two lines, some have one line. iChat has buddy icons in the size of 28×28, which is not a nice size as it has to shrink down a buddy icon usually 48×48 or 50×50 in size. It is also difficult to pick which control to use. Windows Live Messenger for Mac is the only messenger that uses the true list control from OSX, while iChat simply do away with groups for tidiness.

Tabs are also another questioned area in OS X. While iChat in OS X Tiger have tabs at the bottom, iChat in Leopard moved the tabs to the left for the new Source List look, and Safari still have tabs at the top. Yahoo! Messenger followed suit to have tabs at the bottom to be as ‘authentic’ as possible.

Therefore, custom controls are needed no matter what, but the question is how we can design it so it looks like a part of the OS.

Of course, the necessity of custom controls is not the license to go crazy in the design. In order to pay respect to the users of Mac OS X, we paid attention to a lot of details so it can be easier to get used to.

We made numerous small changes in the base version of Trillian Cordonata, the default skin of Trillian. Of two most obvious changes, the window buttons are moved to the position where OS X users are used to. The Zoom effect also maximizes the contact list to its content only. We are also going to make the scroll bar comply with the system setting so the arrow buttons will fall into the right place. The glow and selected states of list items will also correspond to the color settings in System Preferences.

And in fact, concepts from Mac OS X were referenced heavily in the design of Trillian. For example, the Trillian preferences window is based on System Preferences in Mac OS. Even though it has a Windows look in the Windows version currently, the Mac version will spot a Mac look with authentic controls. The design of the skin also has a feel of a ‘remote control’ in mind, similar to the widgets on Dashboard, or the controls of the Apple DVD Player.

In future, we would hopefully allow the possibility of having other interfaces. While I would love to see a ‘normal’ OS X window or a minimal black slate window also, the current plan is to develop Cordonata first, which is essentially the superset of these other alternative skins. Those other skins may be beneficial to those who only uses IM on a Mac.

Next up, I’ll talk about how Trillian Astra for Mac lives well with its brothers on the Web and on Windows.

Proceed to Part 3…

With the release of the alpha of Trillian Astra for Mac, there are lots of comments and controversies stirred up by the new design of the software. In the following 3-part series, I will attempt to explain the reasoning behind the design and open myself to critiques and comments.

iTunes: a straight port

Apple’s Mac OS X is the Mecca of desktop user interface design. Despite how many times Apple itself breaks its own interface guidelines with iTunes, Time Machine or Dashboard, such new experiences are justified by good logic and craftsmanship, no matter how they deviate from the norm.

As we move along in the future of software, I envision that the distinction between downloaded desktop software and web software. And this is how the design challenge began.

Designing between Platforms

Microsoft Office: a culture-specific port

We first saw software from OS X coming to Windows (e.g. iTunes, QuickTime), and software from Windows coming to OS X (e.g. Office), when their user bases were distinct, i.e. Mac users always stay in Mac, Windows users always stay in Windows. That was relatively easy. Windows and Mac OS X could be a metaphor of two different nations with different culture. One could simply translate the software directly into another and call it done, or one could also design the software to be culture-specific. The latter means that, Office for Windows have a Windows layout and Windows features, while Office for Mac have a Mac layout and Mac features.

Adobe Photoshop: an in-between port

And then we also had software that are designed to work on both platforms (e.g. Photoshop). Back in the days, they were mostly professional software. Such software exists as its own entity; their users will be unhappy to relearn the software if they switch OS. Therefore, it is always interesting to look at Adobe and Macromedia products, in the way how they designed the UI to make users of both OS-es happy. In my observation, Adobe is getting more aggressive these days introducing its own UI elements, while still respecting the fonts and metrics of the OS.

And now we have the third type of software: Software that is always connected, across operating systems and across platforms of desktop, web and device.

So here’s the challenge: Design a unified interface across all OS platforms, desktop and web, while retaining an identity. This is Trillian Astra.

Admittedly, Trillian Astra is one of the few first software out there that attempts such triumvirate. Here at Trillian, a lot of users use IM across OS-es and devices. For example, one would have a Mac at home, while a Windows PC in the office, and also an iPhone on the run at the same time. There would also be users who have a Windows PC at home, and always on the run in libraries and Internet cafes. The user does not stay at the same platform all the time in these cases, as we encourage them to use Trillian across the board. In this point of view, coherence is not simply between the software and the operating system, but also between the software in these different incarnations.

I would like to start a conversation with any UX designers out there, and I hope we will have a better reach to do this with the release of the Mac edition. Please feel free to drop your comments in.

Next up, I will talk about how Trillian is designed to be coherent with Mac OS X.

Proceed to Part 2…

Phew! The week was quick! The idea of making Trillian work on the iPhone did not come around until two weeks ago, when we are convinced that Trillian Astra may be able to run on the phone.

There are a few engineering challenges concerning the connectivity of a mobile device, and design limits due to the size and processing power of the device. Since the iPhone is designed as a gesamtkunswerk (a total work of art), my main concern is to create a user experience that is as coherent with the iPhone itself as possible, while still retaining a brand identity: something that won’t make an iPhone user feel like they are contaminating their dear phone, while keeping Trillian users proud.

Here are some quick things that just come out of my brain in regards of designing applications on iPhone:

1. Ergonomics: After dissecting the iPhone UI, I concluded a very simple rule: Buttons need to be the size of a finger. And it cannot be too close to each other.
2. Metrics: Apple keep sizes of things fairly straightforward, everything is around multiples of 10 pixels, which translates to 1/16 of an inch. A button is usually 40 pixels tall, which is a quarter of an inch.
3. Color: The iPhone interface, in fact, uses very similar hue ranges as Trillian Astra. But then all these are just guessing unless we can actually take screenshots from the phone.
4. Interface Inconsistencies: While reading from the iPhone applications, I realize that their interface is actually not very consistent: Sometimes the bottom is black, sometimes is blue, sometimes it’s used as a status bar, sometimes a toolbar. Some screens are black, some screens are white. Nevertheless, it is still pretty tied together because it’s intuitive and minimal enough.
5. Font: Apple loves Helvetica on the iPhone. Humanist fonts seem to go out of style when Microsoft begins adopting them in Windows Vista. Surprisingly enough, the phone includes Helvetica’s archnemesis, Arial, probably for viewing Google, who is obsessed in the font.
6. CSS Compatibility: Most of the time was spent on getting the perfect CSS. I coded all the HTML using just <DIV>’s for reusability. It was a great experience to code a web page that works for just one browser. Makes you wonder why we should all suffer from browser wars. Safari on iPhone also spots some CSS3 features that other browsers don’t have, and I can use them happily with no worries.
7. Testing: You need to have an iPhone to test no matter what. Safari for Windows or Mac may help you verify your stylesheets, but it will not reflect the exact rendering on the iPhone and the way user interacts with the application.

With that said, I have yet to get an iPhone… I know, it’s harder to design if I cannot immerse myself into ‘knowing’ the device I am designing on.

This is ecstatic!!

Can you tell which one is in vector and which one is in bitmap? Cordonata is designed in Adobe Illustrator from ground up, so the conversion has been easy. Scale9Grid feature of Flash 8+ played a pivotal role in composing these windows. It took a week or two of experiment, however, to get bitmaps to scale properly in Flash. Since Flash does not support blurring and drop shadows without turning shapes into movieclips, I used a lot of the ancient “Soften Fill Edges” feature to get my way. The feature turns out to be incredibly effective and shaved lots of KBs off the window, ensuring a fast load for our dear new web users.

Although it is internally codenamed “Express”, it is not a separate product unlike, e.g. AIM vs AIM Express, but instead a subset of the grand vision umbrella-ed under Trillian Astra.

For more information and signing up for testing, the brand new Trillian Astra web site says it all!

…and that was how I described it back then in 2001. The windows are sculpted in a specific shape so that they would line up and create series of choreographed waves going up and down, like rows of mountains… hence the name, “Cordillera”.

I had to design the most versatile, sensible and attractive interface. There was a month’s time to finish the skin, in order to enter the skinning contest. The design underwent a long scrutiny until the final draft. It was not an easy design:

1. All windows need to be resizable, i.e. it has to look in virtually any size – or at least normal sizes – it is unusual in the realm of design.
2. You can have any amount of windows, and that means it has to look good whether there are sixty flowers or one flower in the vase. Both the design of a flower and the vase are equally important and difficult. The Contact List (or “Main Window” as we called it back then) is the only unique and always visible element, while you can have a variety of Message Windows and Status Windows, and so on.
3. The design challenges above still exist today, but to make more difficult, back then I had limited technology to work with. I could not stretch or do fancy calculations (now you can!). I could only tile. I could not overlay things on top of one and another. Text boxes, contact list areas and display areas must have opaque backgrounds. You couldn’t align things at center. Oh, and you couldn’t change any fonts either.

After spending a month of summer vacation working on it, the result was a curvy skin in cerulean blue called “Trillian Professional 1.0“. It was very well-received, and eventually chosen to be the default skin for the then-future version of Trillian “0.64″. It was a great feeling back then to have my work recognized, and also the first time working with strangers over the Internet. I had won a few awards back in high school on designing computer programs for education purposes, but then nothing really happened in college ever since until a few years, so that came as a little surprise. It was pretty exciting back then.

Months of refinement and collaboration led to a total redraw of the skin, soon widely known as “Trillian Cordillera”. Along with the release of a brand new 0.70 version of Trillian that featured file transfer, the software was propelled into stardom within days, gaining momentum and attention from media and enemies alike. Trillian became very popular and eventually its image landed on the cover of PC Magazine and the pages of Wired Magazine and more.

Of course, success was not without price. A few months later, while Trillian was blocked by AOL, Cordillera was also ripped and copied by other skinners in the forum and from other softwares. It took me very long to accept such harsh reality, and realized that it was actually no threat to me since rippers and imitators are always people who are much less talented. But of course, I didn’t know that, and instead, I wasted too much time on defending the skin being ripped off, and eventually became uninterested.

That also was probably the dark ages of the world of Trillian skinning. To push for a professional image, we decided to use Whistler as the default skin for Trillian Pro 1.0, since the other choice “Trillian Cordonata” (a totally different Cordonata than the one you see now) was not good enough. And so we neglected the fluid forms of Cordillera, and were stuck with the corporate rectangle for good four years.

During that period of time, I had designed “Trillian Cordillera 03″ and “Trillian Cordillera 2004″, but neither of them got released, because it required a lot of time and it had limited audience. I also designed for a lot of things that required better skinning technology that I always expected to be included in the next version but they never made it. And so, we were stuck in the middle ages with Whistler.

Trillian Astra project changed everything. With the introduction of Profile+Widgets, Trillian Astra is geared more towards a crowd who likes to express themselves, and as a result, a more expressive skin is required in order to attract such a crowd. The versatile and elegant design of “Trillian Cordonata” made its debut, with a nice balance between the pragmatic Whistler and joyful Cordillera. But the nicer thing is, developing Cordonata results a great by-product: better skinning technologies that we had been looking for since 2001.

So here we are. With much better vector rendering skills and expertise, and with the use of Trillian Astra’s new technologies for alpha-transparent windows, bitmap stretching, PNG layering, skinnable contact list area, color themes and a lot more, let’s welcome the new “Trillian Cordillera Astreme”, a total redraw, rethinking and redesign, a triumphant rebirth of the classic. Though she had now surrendered its crown as the default skin to Cordonata, she remains a shining jewel that we will love. And I hope you’ll like it too.

What’s more, it is not just a screenshot. It is available for download already. Of course, you need to be a tester in order to use it. Enjoy!

Have a very Merry Christmas to everyone!!!